Open-World Multi-Task Control Through Goal-Aware Representation Learning and Adaptive Horizon Prediction

Preparation

Our codebase require Python ≥ 3.9. It also requires a modified version of MineDojo as the simulator and MineCLIP as the goal text encoder. Please run the following commands to prepare the environments.

conda create -n controller python=3.9 
conda activate controller
python -m pip install numpy torch==2.0.0.dev20230208+cu117 --index-url https://download.pytorch.org/whl/nightly/cu117
python -m pip install -r requirements.txt
python -m pip install git+https://github.com/MineDojo/MineCLIP
python -m pip install git+https://github.com/CraftJarvis/MC-Simulator.git

Dataset

Biome	Download
Plains	url
Flat	to be uploaded
Forests	to be uploaded

Train agent

Run the following commands to train the agent.

python main.py data=multi_plains eval=multi_plains

We have provided three configure files for three biomes (multi_plains, multi_forests, and multi_flat).

Running agent models

To run the code, call

python main.py eval=multi_plains eval.only=True model.load_ckpt_path=<path/to/ckpt>

After loading, you should see som windows where agents are playing Minecraft.

Model checkpoints

Below are the configures and weights of models.

Configure	Download	Biome	Number of goals
Transformer	here	Plains	4
Transformer + Extra Observation	here	Plains	4

For example, if we want to use the "Transformer+Extra Observation" checkpoint, we should specify model=transformer_w_extra in the command.

python main.py eval=multi_plains eval.only=True model=transformer_w_extra model.load_ckpt_path=<path/to/ckpt>

Paper and Citation

Our paper is posted on Arxiv. If it helps you, please consider citing us!

@article{cai2023open,
  title={Open-World Multi-Task Control Through Goal-Aware Representation Learning and Adaptive Horizon Prediction},
  author={Cai, Shaofei and Wang, Zihao and Ma, Xiaojian and Liu, Anji and Liang, Yitao},
  journal={arXiv preprint arXiv:2301.10034},
  year={2023}
}

CraftJarvis/MC-Controller